How can I manage Persistent Volumes during a cluster upgrade?

It really boils down to your specific use case for the EBS volumes. If those volumes are tied to critical applications or stateful services, ensure you have robust disaster recovery measures in place.

Yes, blue-green upgrades can be practical if you're behind on versions and need the safety net to roll back if something goes wrong. However, handling storage can get tricky, and it depends on how those persistent volumes are integrated into your workloads. If you're using stateful services, like databases, consider options for replication to minimize risks. Every upgrade should already incorporate a solid disaster recovery plan.

Handling EBS volumes during a blue-green upgrade can depend heavily on your workloads. Ideally, your workload should be designed for high availability, ensuring that a pod's downtime doesn't compromise reliability. Upgrading EKS has always been smooth for us, and we haven't opted for blue-green upgrades before, so it might be worth considering how critical your EBS workloads truly are.

If by blue-green you mean creating an entirely new cluster and decommissioning the old one, that might complicate things unnecessarily. While it can be a one-off situation, managing multiple clusters like that isn’t the standard approach. It's better to assess whether you really need such a complex setup, or if there’s a simpler solution to upgrade components incrementally.